Shielding is widely used to enforce safety in reinforcement learning (RL), ensuring that an agent's actions remain compliant with formal specifications. Classical shielding approaches, however, are often static, in the sense that they assume fixed logical specifications and hand-crafted abstractions. While these static shields provide safety under nominal assumptions, they fail to adapt when environment assumptions are violated. In this paper, we develop the first adaptive shielding framework - to the best of our knowledge - based on Generalized Reactivity of rank 1 (GR(1)) specifications, a tractable and expressive fragment of Linear Temporal Logic (LTL) that captures both safety and liveness properties. Our method detects environment assumption violations at runtime and employs Inductive Logic Programming (ILP) to automatically repair GR(1) specifications online, in a systematic and interpretable way. This ensures that the shield evolves gracefully, ensuring liveness is achievable and weakening goals only when necessary. We consider two case studies: Minepump and Atari Seaquest; showing that (i) static symbolic controllers are often severely suboptimal when optimizing for auxiliary rewards, and (ii) RL agents equipped with our adaptive shield maintain near-optimal reward and perfect logical compliance compared with static shields.

We present a method for computing free-energy differences using thermodynamic integration with a neural network potential that interpolates between two target Hamiltonians. The interpolation is defined at the sample distribution level, and the neural network potential is optimized to match the corresponding equilibrium potential at every intermediate time-step. Once the interpolating potentials and samples are well-aligned, the free-energy difference can be estimated using (neural) thermodynamic integration. To target molecular systems, we simultaneously couple Lennard-Jones and electrostatic interactions and model the rigid-body rotation of molecules. We report accurate results for several benchmark systems: a Lennard-Jones particle in a Lennard-Jones fluid, as well as the insertion of both water and methane solutes in a water solvent at atomistic resolution using a simple three-body neural-network potential.

Today marks exactly 10 years since NASA's Curiosity rover touched down on Mars. The one-tonne vehicle launched from Earth in November 2011 and – after an arduous nine-month journey which included the'seven minutes of terror' down to the Martian surface – it set out to look for evidence that the Red Planet may once have supported life. Since then, Curiosity has driven nearly 18 miles (29 kilometres) and ascended 2,050 feet (625 metres) as it explores Gale Crater and the foothills of Mount Sharp within it. The rover has analysed 41 rock and soil samples, relying on a suite of science instruments to learn what they reveal about Earth's rocky sibling. Such has been its success, what was originally intended to be a two-year mission was later extended indefinitely, leading to a rather busy decade.

Scientists used torpedo-shaped robots to map the Arctic seafloor with sonar, revealing massive sinkholes of thawed permafrost. This story was originally published by Wired and is reproduced here as part of the Climate Desk collaboration. Around 20,000 years ago, the world was so frigid that massive glaciers sucked up enough water to lower sea levels by 400 feet. As the sea pulled back, newly exposed land froze to form permafrost, a mixture of earth and ice that today sprawls across the far north. But as the world warmed into the climate we enjoy today (for the time being), sea levels rose again, submerging the coastal edges of that permafrost, which remain frozen below the water. It's a huge, hidden climate variable that scientists are racing to understand.

Hydropower has been stirring up controversies since the early 2000s. Despite being promoted as a solution to mitigate climate change, the hydropower bubble burst when researchers discovered in 2005 that hydropower dams are responsible for huge amounts of greenhouse gas emissions. Hydropower dams' walls restrict the flow of rivers and turn them into pools of stagnant water. Reservoir surfaces and turbines then release methane into the atmosphere. Methane makes up approximately 80 percent of the greenhouse gases emitted from hydropower dams, peaking in the first decade of the dams lifecycle.

Back in 1950, Enrico Fermi posed the question now known as the Fermi Paradox: Given the countless galaxies, stars, and planets out there, the odds are that life exists elsewhere--so why haven't we found it? The size of the universe is only one possible answer. Maybe humans have already encountered extraterrestrial (ET) life but didn't recognize it. Maybe it doesn't want to be found. Maybe it doesn't find us interesting.

Well, the battery won't allow you to drive for a million miles without recharging, but it will last for a million miles before it must be replaced. This is a big step forward considering EV batteries typically last 200,000 miles. With a million-mile battery, the car would fall apart long before the battery goes dead. This also means the owner can sell it or transfer it to a new car, resulting in less pollution and waste. The brains at Huawei are working on a solution.

With the current year coming to an end, the definition of how businesses leverage technology has changed much due to the pandemic. With disruptive technologies driving global discussion, sustainability is emerging as a new investment. Business leaders are now looking to run their companies in an environmentally sustainable manner, so less harm is done on the planet. Therefore there is a growing emphasis on how technology can be employed for improving a company's environmental performance and the bottom line. From incorporating sustainable practices into business operations to encouraging consumers, employees to embrace sustainability to using AI and quantum computing to find alternate energy-efficient fuels, most of the top enterprises are already doing their part to ensure a greener future.

Polymeric membranes assist with a wide variety of tasks, including water filtration and gas-vapor separation. Designing a membrane for the desired function is more time-consuming than people may expect. However, researchers at Columbia Engineering, Germany's Max Planck Society and the University of South Carolina applied data science to the task to streamline their efforts. More specifically, they combined big data with machine learning to strategically design polymer membranes to act as gas filters. People frequently depend on plastic films and membranes to separate mixtures of simple cases, such as carbon dioxide and methane.

Climate change is seen as the biggest challenge to the future of human life on Earth, and understanding the scientific language used to describe it can sometimes feel just as difficult. But help is at hand. Use our translator tool to find out what some of the words and phrases relating to climate change mean. Keeping the rise in global average temperature below 1.5 degrees Celsius will avoid the worst impacts of climate change, scientists say.